ACCT 303, Fall 2021

Cost Estimation Project Case Submission

2

Question 2: Copy and paste the graph of COGS or SGA on Sales for the years 2005-2018 into this

document for your group's assigned company. Groups of 2 may choose 1 graph (either COGS or SGA

and answer all remaining questions using the same variable.

Groups of 3 must post both graphs

(20

points)

For full credit, your graph must have(a) a descriptive title, (b) axis titles and (c) a linear

trendline with both the regression equation and R-square.

Based on the graph, does it look like a linear relationship exists between that variable and

Sales.

Question 3: Copy and paste the simple regression

of COGS or SGA on Sales for the years 2005-2018 into

this document for your group's assigned company. Groups of 2 may choose 1 regression (either COGS or

SGA and answer all remaining questions using the same variable.

Groups of 3 must post both regressions

.

Is the regression reliable?

Comment by using the following criteria in the from Table S.7 from your text's

regression supplement (in Moodle) or page 4 on case instruction handout:

(25 points)

R-square (for goodness of fit of the overall regression, if 75% or better);

F-statistic for statistical reliability (associated significance level should be less than 0.05);

t-statistic for reliability of each individual variable

(generally each variable should have a t-

statistic over 2 or its associate p-value of less than 0.05).

Question 3: Copy and paste the multiple regression

of COGS or SGA

and on Sales AND Sales_NG for

the years 2005-2018 into this document for your group's assigned company. Groups of 2 may choose 1

regression (either COGS or SGA and answer all remaining questions using the same variable.

Groups of

3 must post both regressions

.

Is the regression reliable?

Comment by using the following criteria in the

from Table S.7 from your text's regression supplement (in Moodle) or page 4 on case instruction handout:

(25 points)

R-square (for goodness of fit of the overall regression, if 75% or better);

F-statistic for statistical reliability (associated significance level should be less than 0.05);

t-statistic for reliability of each individual variable (generally each variable should have a t-

statistic over 2 or its associate p-value of less than 0.05).

Question 4: Based on the results of the multiple regression the costs for your assigned company To

diagnose whether costs are "sticky".

(5 points)

.

To be sticky, the coefficient on Sales_NG from the multiple

regression must both (a) positive and (b) statistically significant with a t-value greater than 2 and associated p-

value less than five percent (0.05).

If the coefficient on Sales_NG is positive with a t-value above 2, then one

would multiple Sales times a higher coefficient (sum of coefficient on Sales plus coefficient on Sales_NG) in

years where sales decreased.

In other words

,

COGS is estimated to be a higher percent of sales in years when

sales decrease than in years when sales stay the same or increase.

ACCT 303, Fall 2021

Cost Estimation Project Case Submission

3

Question 5: How does the multiple regression (on Sales and Sales_NG) perform in predicting 2019 COGS

and SGA versus the simple regression

(5 points).

If you have pasted the regression results into the correct

rows (N31 for simple regression of COGS, N64 for multiple regression of COGS, Y31 for simple regression of

SGA, Y64 for multiple regression of SGA), the predictions and the prediction error for each regression is

calculated for you (in row 55 for the simple regressions and row 89 for the multiple regressions

).

If the

absolute value of the error for the multiple regression is less than the absolute value of the error for the simple

regression, the multiple regression performed better.

Groups of 2 may answer this question using either COGS

or SGA, using the same variable chosen to answer questions 2 to 4.

Groups of three must answer this question

for both COGS and SGA.

Note the remaining 5 points are from each student filling answering all questions in a short survey in

Moodle, reflecting on the cost estimation case